博主:handsomecui,希望路过的各位大佬留下你们宝贵的意见,在这里祝大家冬至快乐。
缘由:
初探缘由,在业务层想要通过(当前线程的栈)来获取到控制层的类名,然后打日志,可是发现并不能通过当前线程获取到控制层的类,两者并没有在一个线程内,进而引发了我对这一问题的思考。
解决步骤:
一. 我先打印出了当前运行的所有线程的名字以及栈的运行情况:
Thread list size == 21
线程名:main
java.net.PlainSocketImpl:socketAccept
java.net.AbstractPlainSocketImpl:accept
java.net.ServerSocket:implAccept
java.net.ServerSocket:accept
org.apache.catalina.core.StandardServer:await
org.apache.catalina.startup.Catalina:await
org.apache.catalina.startup.Catalina:start
sun.reflect.NativeMethodAccessorImpl:invoke0
sun.reflect.NativeMethodAccessorImpl:invoke
sun.reflect.DelegatingMethodAccessorImpl:invoke
java.lang.reflect.Method:invoke
org.apache.catalina.startup.Bootstrap:start
org.apache.catalina.startup.Bootstrap:main
线程名:ContainerBackgroundProcessor[StandardEngine[Catalina]]
java.io.UnixFileSystem:list
java.io.File:list
org.apache.catalina.startup.HostConfig:deployApps
org.apache.catalina.startup.HostConfig:check
org.apache.catalina.startup.HostConfig:lifecycleEvent
org.apache.catalina.util.LifecycleSupport:fireLifecycleEvent
org.apache.catalina.util.LifecycleBase:fireLifecycleEvent
org.apache.catalina.core.ContainerBase:backgroundProcess
org.apache.catalina.core.ContainerBase$ContainerBackgroundProcessor:processChildren
org.apache.catalina.core.ContainerBase$ContainerBackgroundProcessor:processChildren
org.apache.catalina.core.ContainerBase$ContainerBackgroundProcessor:run
java.lang.Thread:run
线程名:http-bio-8080-Acceptor-0
java.net.PlainSocketImpl:socketAccept
java.net.AbstractPlainSocketImpl:accept
java.net.ServerSocket:implAccept
java.net.ServerSocket:accept
org.apache.tomcat.util.net.DefaultServerSocketFactory:acceptSocket
org.apache.tomcat.util.net.JIoEndpoint$Acceptor:run
java.lang.Thread:run
线程名:http-bio-8080-AsyncTimeout
java.lang.Thread:sleep
org.apache.tomcat.util.net.JIoEndpoint$AsyncTimeout:run
java.lang.Thread:run
线程名:ajp-bio-8009-Acceptor-0
java.net.PlainSocketImpl:socketAccept
java.net.AbstractPlainSocketImpl:accept
java.net.ServerSocket:implAccept
java.net.ServerSocket:accept
org.apache.tomcat.util.net.DefaultServerSocketFactory:acceptSocket
org.apache.tomcat.util.net.JIoEndpoint$Acceptor:run
java.lang.Thread:run
线程名:ajp-bio-8009-AsyncTimeout
java.lang.Thread:sleep
org.apache.tomcat.util.net.JIoEndpoint$AsyncTimeout:run
java.lang.Thread:run
线程名:http-bio-8080-exec-1
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:park
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject:await
java.util.concurrent.LinkedBlockingQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
org.apache.tomcat.util.threads.TaskThread$WrappingRunnable:run
java.lang.Thread:run
线程名:http-bio-8080-exec-2
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:park
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject:await
java.util.concurrent.LinkedBlockingQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
org.apache.tomcat.util.threads.TaskThread$WrappingRunnable:run
java.lang.Thread:run
线程名:http-bio-8080-exec-3
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:park
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject:await
java.util.concurrent.LinkedBlockingQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
org.apache.tomcat.util.threads.TaskThread$WrappingRunnable:run
java.lang.Thread:run
线程名:http-bio-8080-exec-4
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:park
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject:await
java.util.concurrent.LinkedBlockingQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
org.apache.tomcat.util.threads.TaskThread$WrappingRunnable:run
java.lang.Thread:run
线程名:http-bio-8080-exec-5
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:park
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject:await
java.util.concurrent.LinkedBlockingQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
org.apache.tomcat.util.threads.TaskThread$WrappingRunnable:run
java.lang.Thread:run
线程名:http-bio-8080-exec-6
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:park
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject:await
java.util.concurrent.LinkedBlockingQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
org.apache.tomcat.util.threads.TaskThread$WrappingRunnable:run
java.lang.Thread:run
线程名:http-bio-8080-exec-7
java.lang.Thread:getStackTrace
com.thor.dao.domain.common.util.CommonUtil:getThreadGroup
com.thor.service.customer.CustomerService:post
com.thor.service.customer.CustomerService$$FastClassByCGLIB$$9a694ea:invoke
net.sf.cglib.proxy.MethodProxy:invoke
org.springframework.aop.framework.Cglib2AopProxy$CglibMethodInvocation:invokeJoinpoint
org.springframework.aop.framework.ReflectiveMethodInvocation:proceed
org.springframework.transaction.interceptor.TransactionInterceptor:invoke
org.springframework.aop.framework.ReflectiveMethodInvocation:proceed
org.springframework.aop.framework.Cglib2AopProxy$DynamicAdvisedInterceptor:intercept
com.thor.service.customer.CustomerService$$EnhancerByCGLIB$$898851cc:post
com.thor.control.customer.CustomControl:post
sun.reflect.NativeMethodAccessorImpl:invoke0
sun.reflect.NativeMethodAccessorImpl:invoke
sun.reflect.DelegatingMethodAccessorImpl:invoke
java.lang.reflect.Method:invoke
org.springframework.web.method.support.InvocableHandlerMethod:invoke
org.springframework.web.method.support.InvocableHandlerMethod:invokeForRequest
org.springframework.web.servlet.mvc.method.annotation.ServletInvocableHandlerMethod:invokeAndHandle
org.springframework.web.servlet.mvc.method.annotation.RequestMappingHandlerAdapter:invokeHandlerMethod
org.springframework.web.servlet.mvc.method.annotation.RequestMappingHandlerAdapter:handleInternal
org.springframework.web.servlet.mvc.method.AbstractHandlerMethodAdapter:handle
org.springframework.web.servlet.DispatcherServlet:doDispatch
org.springframework.web.servlet.DispatcherServlet:doService
org.springframework.web.servlet.FrameworkServlet:processRequest
org.springframework.web.servlet.FrameworkServlet:doPost
javax.servlet.http.HttpServlet:service
javax.servlet.http.HttpServlet:service
org.apache.catalina.core.ApplicationFilterChain:internalDoFilter
org.apache.catalina.core.ApplicationFilterChain:doFilter
org.apache.tomcat.websocket.server.WsFilter:doFilter
org.apache.catalina.core.ApplicationFilterChain:internalDoFilter
org.apache.catalina.core.ApplicationFilterChain:doFilter
com.thor.control.user.web.filter.AuthFilter:doFilterInternal
com.taovip.login.session.auth.BaseFilter:doFilter
org.apache.catalina.core.ApplicationFilterChain:internalDoFilter
org.apache.catalina.core.ApplicationFilterChain:doFilter
com.thor.control.user.web.filter.LogFilter:doFilter
org.apache.catalina.core.ApplicationFilterChain:internalDoFilter
org.apache.catalina.core.ApplicationFilterChain:doFilter
com.taovip.login.session.client.ClientSessionFilter:doFilter
org.apache.catalina.core.ApplicationFilterChain:internalDoFilter
org.apache.catalina.core.ApplicationFilterChain:doFilter
com.thor.control.user.web.filter.LoginFilter:doFilter
org.apache.catalina.core.ApplicationFilterChain:internalDoFilter
org.apache.catalina.core.ApplicationFilterChain:doFilter
org.springframework.web.filter.CharacterEncodingFilter:doFilterInternal
org.springframework.web.filter.OncePerRequestFilter:doFilter
org.apache.catalina.core.ApplicationFilterChain:internalDoFilter
org.apache.catalina.core.ApplicationFilterChain:doFilter
org.apache.catalina.core.StandardWrapperValve:invoke
org.apache.catalina.core.StandardContextValve:invoke
org.apache.catalina.authenticator.AuthenticatorBase:invoke
org.apache.catalina.core.StandardHostValve:invoke
org.apache.catalina.valves.ErrorReportValve:invoke
org.apache.catalina.valves.AccessLogValve:invoke
org.apache.catalina.core.StandardEngineValve:invoke
org.apache.catalina.connector.CoyoteAdapter:service
org.apache.coyote.http11.AbstractHttp11Processor:process
org.apache.coyote.AbstractProtocol$AbstractConnectionHandler:process
org.apache.tomcat.util.net.JIoEndpoint$SocketProcessor:run
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
org.apache.tomcat.util.threads.TaskThread$WrappingRunnable:run
java.lang.Thread:run
线程名:http-bio-8080-exec-8
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:park
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject:await
java.util.concurrent.LinkedBlockingQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
org.apache.tomcat.util.threads.TaskThread$WrappingRunnable:run
java.lang.Thread:run
线程名:http-bio-8080-exec-9
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:park
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject:await
java.util.concurrent.LinkedBlockingQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
org.apache.tomcat.util.threads.TaskThread$WrappingRunnable:run
java.lang.Thread:run
线程名:http-bio-8080-exec-10
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:park
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject:await
java.util.concurrent.LinkedBlockingQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
org.apache.tomcat.util.threads.TaskQueue:take
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
org.apache.tomcat.util.threads.TaskThread$WrappingRunnable:run
java.lang.Thread:run
线程名:mergeAppender
java.lang.Object:wait
java.lang.Object:wait
org.apache.log4j.AsyncAppender$Dispatcher:run
java.lang.Thread:run
线程名:alertAppender
java.lang.Object:wait
java.lang.Object:wait
org.apache.log4j.AsyncAppender$Dispatcher:run
java.lang.Thread:run
线程名:Timer-0
java.lang.Object:wait
java.util.TimerThread:mainLoop
java.util.TimerThread:run
线程名:pool-7-thread-1
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:parkNanos
java.util.concurrent.SynchronousQueue$TransferStack:awaitFulfill
java.util.concurrent.SynchronousQueue$TransferStack:transfer
java.util.concurrent.SynchronousQueue:poll
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
java.lang.Thread:run
线程名:pool-7-thread-2
sun.misc.Unsafe:park
java.util.concurrent.locks.LockSupport:parkNanos
java.util.concurrent.SynchronousQueue$TransferStack:awaitFulfill
java.util.concurrent.SynchronousQueue$TransferStack:transfer
java.util.concurrent.SynchronousQueue:poll
java.util.concurrent.ThreadPoolExecutor:getTask
java.util.concurrent.ThreadPoolExecutor:runWorker
java.util.concurrent.ThreadPoolExecutor$Worker:run
java.lang.Thread:run
代码:
ThreadGroup threadGroup = Thread.currentThread().getThreadGroup();
int estimatedSize = threadGroup.activeCount() * 2;
Thread[] slackList = new Thread[estimatedSize];
// 获取根线程组的所有线程
int actualSize = threadGroup.enumerate(slackList);
// copy into a list that is the exact size
Thread[] list = new Thread[actualSize];
System.arraycopy(slackList, 0, list, 0, actualSize);
System.out.println("Thread list size == " + list.length);
for (Thread thread : list) {
System.out.println("线程名:" + thread.getName());
StackTraceElement[] stackTrace = thread.getStackTrace();
for(StackTraceElement e : stackTrace){
System.out.println(e.getClassName()+":"+e.getMethodName());
}
}
所以看到了这么多线程,网上搜了好多,都看不懂,终于找到了一篇只讲名功能的博客:https://www.cnblogs.com/dhcn/p/7120713.html
Work线程
功能
HTTP请求的处理线程(非NIO)。当有新的http请求进来后,则会从线程池中获得一个线程Work对象,调用Work.assign函数,将新到的http请求分配给这个线程。
名称
名称是http-[IpAddr]-[Port]-[Number],如http-0.0.0.0-8080-1
这个可以从Http11Protocol中的setName函数和Worker中的start方法得知这个命名方式。
1 | public String getName() { |
2 | String encodedAddr = ""; |
3 | if (getAddress() != null) { |
4 | encodedAddr = "" + getAddress(); |
5 | if (encodedAddr.startsWith("/" )) |
6 | encodedAddr = encodedAddr.substring(1); |
7 | encodedAddr = URLEncoder. encode(encodedAddr) + "-"; |
8 | } |
9 | |
10 | return ("http-" + encodedAddr + endpoint.getPort()); |
11 | } |
12 | |
13 |
线程类:JIoEndpoint.Work
在JIoEndpoint.Work的run方法中调用await方法等待并获得下一个socket,传给handle进行处理。在await方法中,如果没有分配新的客户端请求socket, available变量会一直false,并会循环调用wait方法阻塞自己,同时释放Work对象的锁,直到Acceptor线程获得新的socket, 并调用Work.assign方法分配给该工作线程。 这时availble变量才为设置为true,并且await方法会返回分配的socket对象。
1 | protected class Worker implements Runnable { |
2 | |
3 | protected Thread thread = null; |
4 | |
5 | protected boolean available = false; |
6 | |
7 | protected Socket socket = null; |
8 | |
9 | /** |
10 | |
11 | * Process an incoming TCP/IP connection on the specified socket. Any |
12 | |
13 | * exception that occurs during processing must be logged and swallowed. |
14 | |
15 | * <b>NOTE</b> : This method is called from our Connector's thread. We |
16 | |
17 | * must assign it to our own thread so that multiple simultaneous |
18 | |
19 | * requests can be handled. |
20 | |
21 | * |
22 | |
23 | * @param socket TCP socket to process |
24 | |
25 | */ |
26 | |
27 | synchronized void assign(Socket socket ) { |
28 | |
29 | // Wait for the Processor to get the previous Socket |
30 | |
31 | while (available ) { |
32 | |
33 | try { |
34 | |
35 | wait(); |
36 | |
37 | } catch (InterruptedException e) { |
38 | |
39 | } |
40 | |
41 | } |
42 | |
43 | // Store the newly available Socket and notify our thread |
44 | |
45 | this.socket = socket ; |
46 | |
47 | available = true ; |
48 | |
49 | notifyAll(); |
50 | |
51 | } |
52 | |
53 | /** |
54 | |
55 | * 等待新分配的Socket |
56 | |
57 | */ |
58 | |
59 | private synchronized Socket await() { |
60 | |
61 | //等待Connector提供新的Socket |
62 | |
63 | while (!available ) { |
64 | |
65 | try { |
66 | |
67 | wait(); |
68 | |
69 | } catch (InterruptedException e) { |
70 | |
71 | } |
72 | |
73 | } |
74 | |
75 | //通知Connector我们已经接收到这个Socket |
76 | |
77 | Socket socket = this.socket ; |
78 | |
79 | available = false ; |
80 | |
81 | notifyAll(); |
82 | |
83 | return (socket); |
84 | |
85 | } |
86 | |
87 | /** |
88 | |
89 | * 后台线程,监听进入的TCP/IP连接,并传递给合适的处理模块 |
90 | |
91 | */ |
92 | |
93 | public void run() { |
94 | |
95 | // Process requests until we receive a shutdown signal |
96 | |
97 | //处理请求直到我们接收到shutdown信号 |
98 | |
99 | while (running ) { |
100 | |
101 | //等待下一个分配的socket |
102 | |
103 | Socket socket = await(); |
104 | |
105 | if (socket == null) |
106 | |
107 | continue; |
108 | |
109 | //设置socket的选项,并处理socket |
110 | |
111 | if (!setSocketOptions(socket) || !handler.process(socket)) { |
112 | |
113 | // 关闭socket |
114 | |
115 | try { |
116 | |
117 | socket.close(); |
118 | |
119 | } catch (IOException e) { |
120 | |
121 | } |
122 | |
123 | } |
124 | |
125 | // Finish up this request |
126 | |
127 | socket = null; |
128 | |
129 | //回收线程 |
130 | |
131 | recycleWorkerThread( this); |
132 | |
133 | } |
134 | |
135 | } |
136 | |
137 | /** |
138 | |
139 | * 开启后台处理线程 |
140 | |
141 | */ |
142 | |
143 | public void start() { |
144 | |
145 | thread = new Thread(this); |
146 | |
147 | thread.setName(getName() + "-" + (++curThreads)); |
148 | |
149 | thread.setDaemon(true); |
150 | |
151 | thread.start(); |
152 | |
153 | } |
154 | |
155 | } |
156 | |
157 |
所属线程池
所属线程池实现功能比较简单,是内嵌到JIoEndpoint类中的实现。基本数据结构是一个工作线程栈JIoEndpoint.WorkerStack。
线程池主要属性
curThreadsBusy:当前繁忙线程数
curThreads:当前工作线程数
maxThreads:最大工作线程数
线程池启动
这个线程池实现功能比较简单,不需要太多启动功能。可以从JIoEndpoint类的start方法看到,启动初始化需要做的事是分配线程栈worker空间。
任务分配时序图
任务分配
通过JIoEndPoint中createWorkerThread方法获得一个工作线程。如在工作线程栈workers中获得一个线程对象,如果线程栈已经是空的,并且当前线程数量curThreads还小于最大线程数maxThreads,那么就创建一个新的工作线程。然后调用Work.assign方法分配给工作线程。
1 | protected Worker createWorkerThread() { |
2 | |
3 | //获得工作线程栈workers的锁 |
4 | |
5 | synchronized (workers ) { |
6 | |
7 | //如果工作线程栈里有线程则返回栈顶工作线程 |
8 | |
9 | if (workers .size() > 0) { |
10 | |
11 | curThreadsBusy++; |
12 | |
13 | return workers .pop(); |
14 | |
15 | } |
16 | |
17 | //如果工作线程栈里没有线程,maxThreads大于0且当前线程数小于最大线程数,则创建一个新的线程 |
18 | |
19 | if ((maxThreads > 0) && (curThreads < maxThreads)) { |
20 | |
21 | curThreadsBusy++; |
22 | |
23 | return (newWorkerThread()); |
24 | |
25 | } else { |
26 | |
27 | //如果maxThreads小于0,则说明没有限制,创建新的线程 |
28 | |
29 | if (maxThreads < 0) { |
30 | |
31 | curThreadsBusy++; |
32 | |
33 | return (newWorkerThread()); |
34 | |
35 | } else { |
36 | |
37 | return (null); |
38 | |
39 | } |
40 | |
41 | } |
42 | |
43 | } |
44 | |
45 | } |
46 | |
47 |
工作线程回收
JIoEndPoint中recycleWorkerThread方法是回收工作线程,当http请求处理完成,则调用该方法回收工作线程。该方法首先获得worker对象锁,然后调用workers.push方法将工作线程压入工作线程栈中,接着将当前繁忙线程数减1,最后调用workers.notify方法。
1 | protected void recycleWorkerThread(Worker workerThread) { |
2 | |
3 | synchronized (workers ) { |
4 | |
5 | workers.push(workerThread); |
6 | |
7 | curThreadsBusy--; |
8 | |
9 | workers.notify(); |
10 | |
11 | } |
12 | } |
配置
在Tomcat中配置文件Server.xml中的Connector属性配置最大线程数maxThreads。
例如:
<Connector port="8080"
maxThreads="150"
……/>
Acceptor线程
功能
获得HTTP请求socket。并从工作线程池中获得一个线程,将socket分配给一个工作线程。
名称
http-[IPAddr]-[Port]-Acceptor-[Number],如http-0.0.0.0-8080-Acceptor-1
线程类:JIoEndpoint.Acceptor
所属线程池
无
启动时序图
在启动时会开启Accepter线程,时序图如下:
线程启动
如上时序图,在Tomcat启动过程会调用JIoEndpoint类的start方法,会创建并启动acceptorThreadCount个Acceptor线程。
1 | public void start() throws Exception { |
2 | |
3 | // Initialize socket if not done before |
4 | |
5 | if (!initialized ) { |
6 | |
7 | init(); |
8 | |
9 | } |
10 | |
11 | if (!running ) { |
12 | |
13 | running = true ; |
14 | |
15 | paused = false ; |
16 | |
17 | //如果没有配置executor线程池,则创建工作线程栈worker, 就是上例中的线程池的工作线程栈。 |
18 | |
19 | if (executor == null) { |
20 | |
21 | workers = new WorkerStack(maxThreads); |
22 | |
23 | } |
24 | |
25 | //启动acceptor线程 |
26 | |
27 | for (int i = 0; i < acceptorThreadCount; i++) { |
28 | |
29 | Thread acceptorThread = new Thread(new Acceptor(), getName() + "-Acceptor-" + i); |
30 | |
31 | acceptorThread.setPriority( threadPriority); |
32 | |
33 | acceptorThread.setDaemon( daemon); |
34 | |
35 | acceptorThread.start(); |
36 | |
37 | } |
38 | |
39 | } |
40 | |
41 | } |
属性
acceptorThreadCount:开启的acceptor线程数,从源码看到这个值并没有通过配置设置,而是固定的值为1
配置
无
Main主线程
功能
完成装配、初始化和启动,之后会开启SocketServer,并循环等待命令,如shutdown。
名称:Main
线程类:Main主线程
所属线程池:
无
catalina-exec线程
功能
StandardThreadExecutor的工作线程,功能和Work线程类似。如果为Connector配置了Executor,则会使用该线程处理http请求。
线程类:ThreadPoolExecutor.Work
所属线程池:StandardThreadExecutor
类名是org.apache.catalina.core.StandardThreadExecutor,该线程池类通过代理设计模式对JavaConcurrent包中的线程池ThreadPoolExecutor进行简单的封装。并实现了Lifecycle接口,以及增加了发送消息的功能。
属性
minSpareThreads:最小空闲线程数
maxThreads:最大线程数
maxIdleTime:最大空闲时间
配置
在Server.xml文件中配置Executor节点,支持如下属性,
Name |
Executor的名称 |
namePrefix |
工作线程前缀 |
maxThreads |
最大线程数 |
minSpareThreads |
最小空闲线程数 |
maxIdleTime |
最大空闲时间 |
并在Connector节点配置executor,并指定为Executor的名称。
例如:
<Executor name="tomcatThreadPool" namePrefix="catalina-exec-" maxThreads="150" minSpareThreads="4" maxIdleTime="200"/>
<Connector Address="0.0.0.0" port="8080" protocol="HTTP/1.1"executor="tomcatThreadPool".../>
TP-Processor线程
功能
AJP协议中Servlet容器的处理线程
名称
TP-Processor-[Number],例如TP-Processor-1
线程类:ThreadPool.ControlRunnable
所属线程池:org.apache.tomcat.util.threads.ThreadPool
该线程池还会启动一个TP-Monitor线程监控空闲线程。在TheadPool会有一个ControlRunnable数组保存线程池中的工作线程。使用该线程池需要先调用start方法,进行ControlRunnable数组初始化,minSpareThreads个空闲线程的创建,以及TP-Monitor线程的启动。
属性
maxThreads:最大线程数
minSpareThreads:最小空闲线程数
maxSpareThreads: 最大空闲线程数
线程池的启动
通过ThreadPool.start方法,该方法会分配线程数组pool,并打开minSpareThreads空线程。如果最大空闲线程数小于最大线程数,则启动TP-Monitor线程。
1 | public synchronized void start() { |
2 | |
3 | stopThePool=false ; |
4 | |
5 | currentThreadCount = 0; |
6 | |
7 | currentThreadsBusy = 0; |
8 | |
9 | adjustLimits(); |
10 | |
11 | pool = new ControlRunnable[maxThreads]; |
12 | |
13 | //启动minSpareThreads空闲线程 |
14 | |
15 | openThreads( minSpareThreads); |
16 | |
17 | //如果最大空闲线程数小于最大线程数,则启动TP-Monitor线程 |
18 | |
19 | if (maxSpareThreads < maxThreads) { |
20 | |
21 | monitor = new MonitorRunnable(this); |
22 | |
23 | } |
24 | |
25 | } |
任务分配
使用ThreadPool.runIt来运行新的任务,在该方法中,会调用findControlRunnable方法来获得一个工作线程。需要注意的是调用方不需要调用额外的方法来回收线程。当ControlRunnable线程完成指定的任务会自动将线程回收到线程池中。
findControlRunnable是ThreadPool线程池的关键方法,它提供了从线程池中获得一个工作线程,并将相应的计数调整,如 tpOpen,currentThreadsBusy。
1 | /** |
2 | |
3 | * Executes a given Runnable on a thread in the pool, block if needed. |
4 | |
5 | */ |
6 | |
7 | public void runIt(ThreadPoolRunnable r) { |
8 | |
9 | if(null == r) { |
10 | |
11 | throw new NullPointerException(); |
12 | |
13 | } |
14 | |
15 | //从线程池中获得一个工作线程 |
16 | |
17 | ControlRunnable c = findControlRunnable(); |
18 | |
19 | //运行任务 |
20 | |
21 | c.runIt(r); |
22 | |
23 | } |
24 | |
25 | private ControlRunnable findControlRunnable() { |
26 | |
27 | ControlRunnable c= null; |
28 | |
29 | if ( stopThePool ) { |
30 | |
31 | throw new IllegalStateException(); |
32 | |
33 | } |
34 | |
35 | //从线程池中获得一个空闲线程 |
36 | |
37 | synchronized(this ) { |
38 | |
39 | //当前繁忙线程和当前线程数相同,则表示所有的开启线程都是繁忙的。 |
40 | |
41 | while (currentThreadsBusy == currentThreadCount) { |
42 | |
43 | //如果当前线程数比最大线程数小 |
44 | |
45 | if (currentThreadCount < maxThreads) { |
46 | |
47 | // Not all threads were open, |
48 | |
49 | // Open new threads up to the max number of idel threads |
50 | |
51 | |
52 | int toOpen = currentThreadCount + minSpareThreads; |
53 | |
54 | openThreads(toOpen); |
55 | |
56 | } else { |
57 | |
58 | logFull(log, currentThreadCount, maxThreads ); |
59 | |
60 | //线程数已经满了,等待线程成为空闲线程 |
61 | |
62 | try { |
63 | |
64 | this.wait(); |
65 | |
66 | } |
67 | |
68 | // was just catch Throwable -- but no other |
69 | |
70 | // exceptions can be thrown by wait, right? |
71 | |
72 | // So we catch and ignore this one, since |
73 | |
74 | // it'll never actually happen, since nowhere |
75 | |
76 | // do we say pool.interrupt(). |
77 | |
78 | catch(InterruptedException e) { |
79 | |
80 | log.error("Unexpected exception" , e); |
81 | |
82 | } |
83 | |
84 | if( log .isDebugEnabled() ) { |
85 | |
86 | log.debug("Finished waiting: CTC=" +currentThreadCount + |
87 | |
88 | ", CTB=" + currentThreadsBusy ); |
89 | |
90 | } |
91 | |
92 | // Pool was stopped. Get away of the pool. |
93 | |
94 | if( stopThePool ) { |
95 | |
96 | break; |
97 | |
98 | } |
99 | |
100 | } |
101 | |
102 | } |
103 | |
104 | //线程池已经关闭,离开线程池 |
105 | |
106 | if(0 == currentThreadCount || stopThePool) { |
107 | |
108 | throw new IllegalStateException(); |
109 | |
110 | } |
111 | |
112 | //到了这里,表示有空闲线程可用 |
113 | |
114 | //取出数组pool中最后一个线程 |
115 | |
116 | int pos = currentThreadCount - currentThreadsBusy - 1; |
117 | |
118 | c = pool[pos]; |
119 | |
120 | pool[pos] = null; |
121 | |
122 | //繁忙线程数加1 |
123 | |
124 | currentThreadsBusy++; |
125 | |
126 | } |
127 | |
128 | return c; |
129 | |
130 | } |
131 | |
132 | /** |
133 | |
134 | *开启线程 |
135 | |
136 | * @param toOpen 我们将要开启的线程数 |
137 | |
138 | */ |
139 | |
140 | protected void openThreads(int toOpen) { |
141 | |
142 | if(toOpen > maxThreads ) { |
143 | |
144 | toOpen = maxThreads; |
145 | |
146 | } |
147 | |
148 | //创建空闲线程 |
149 | |
150 | for(int i = currentThreadCount ; i < toOpen ; i++) { |
151 | |
152 | //需要减去currentThreadsBusy, 因为繁忙线程已经从pool数组中移出 |
153 | |
154 | pool[i - currentThreadsBusy ] = new ControlRunnable( this); |
155 | |
156 | } |
157 | |
158 | currentThreadCount = toOpen; |
159 | |
160 | } |
工作线程回收
通过ThreadPool.returnController方法回收线程。该方法会将繁忙线程数currentThreadsBusy减1,并将线程回收到线程数组中。
1 | /** |
2 | |
3 | * 将线程返还线程池 |
4 | |
5 | */ |
6 | protected synchronized void returnController (ControlRunnable c) { |
7 | |
8 | if(0 == currentThreadCount || stopThePool) { |
9 | |
10 | c.terminate(); |
11 | |
12 | return; |
13 | |
14 | } |
15 | |
16 | // atomic |
17 | |
18 | currentThreadsBusy--; |
19 | |
20 | //将线程回收到pool数组中 |
21 | |
22 | pool[currentThreadCount - currentThreadsBusy - 1] = c; |
23 | |
24 | //notify会唤醒在等待线程资源 |
25 | |
26 | notify(); |
27 | |
28 | } |
配置
在Server.xml文件中配置Connector属性
maxThreads |
最大线程数 |
minSpareThreads |
最小空闲线程数 |
maxSpareThreads |
最大空闲线程数 |
例如:
<Connector port="8009" protocol="AJP/1.3" redirectPort="8443" maxThreads="800" minSpareThreads="50" maxSpareThreads="500" />
TP-Monitor线程
功能
监控ThreadPool线程池的空闲线程,回收比最大空闲线程数多出的空闲线程。
线程类:ThreadPool.MonitorRunnable
1 | /** |
2 | |
3 | * 定期清理空闲线程 |
4 | |
5 | */ |
6 | |
7 | public static class MonitorRunnable implements Runnable { |
8 | |
9 | ThreadPool p; |
10 | |
11 | Thread t; |
12 | |
13 | int interval =WORK_WAIT_TIMEOUT; |
14 | |
15 | boolean shouldTerminate ; |
16 | |
17 | MonitorRunnable(ThreadPool p) { |
18 | |
19 | this.p =p; |
20 | |
21 | this.start(); |
22 | |
23 | } |
24 | |
25 | public void start() { |
26 | |
27 | shouldTerminate = false ; |
28 | |
29 | t = new Thread(this); |
30 | |
31 | t.setDaemon( p.getDaemon() ); |
32 | |
33 | t.setName( p.getName() + "-Monitor"); |
34 | |
35 | t.start(); |
36 | |
37 | } |
38 | |
39 | public void setInterval(int i ) { |
40 | |
41 | this.interval =i; |
42 | |
43 | } |
44 | |
45 | public void run() { |
46 | |
47 | while(true ) { |
48 | |
49 | try { |
50 | |
51 | //Wait一段时间 |
52 | |
53 | synchronized(this ) { |
54 | |
55 | this.wait(interval ); |
56 | |
57 | } |
58 | |
59 | // Check if should terminate. |
60 | |
61 | // termination happens when the pool is shutting down. |
62 | |
63 | if(shouldTerminate ) { |
64 | |
65 | break; |
66 | |
67 | } |
68 | |
69 | //回收空闲线程 |
70 | |
71 | p.checkSpareControllers(); |
72 | |
73 | } catch(Throwable t) { |
74 | |
75 | ThreadPool. log.error("Unexpected exception" , t); |
76 | |
77 | } |
78 | |
79 | } |
80 | |
81 | } |
82 | |
83 | public void stop() { |
84 | |
85 | this.terminate(); |
86 | |
87 | } |
88 | |
89 | /** 停止monitor线程 |
90 | |
91 | */ |
92 | |
93 | public synchronized void terminate() { |
94 | |
95 | shouldTerminate = true ; |
96 | |
97 | this.notify(); |
98 | |
99 | } |
100 | |
101 | } |
ThreadPool.checkSpareControllers方法,用来被TP-Monitor线程调用回收工作线程。
1 | /** |
2 | |
3 | * 被TP-Monitor线程用来回收线程 |
4 | |
5 | */ |
6 | |
7 | protected synchronized void checkSpareControllers() { |
8 | |
9 | if(stopThePool ) { |
10 | |
11 | return; |
12 | |
13 | } |
14 | |
15 | //如果当前空闲线程数大于最大空闲线程数 |
16 | |
17 | if((currentThreadCount - currentThreadsBusy) > maxSpareThreads) { |
18 | |
19 | //回收比最大空闲线程数多出的空闲线程 |
20 | |
21 | int toFree = currentThreadCount - |
22 | |
23 | currentThreadsBusy - |
24 | |
25 | maxSpareThreads; |
26 | |
27 | for(int i = 0 ; i < toFree ; i++) { |
28 | |
29 | ControlRunnable c = pool[currentThreadCount - currentThreadsBusy - 1]; |
30 | |
31 | c.terminate(); |
32 | |
33 | pool[currentThreadCount - currentThreadsBusy - 1] = null; |
34 | |
35 | currentThreadCount --; |
36 | |
37 | } |
38 | |
39 | } |
40 | |
41 | } |
所属线程池
ThreadPool线程池
ContainerBackgroundProcessor线程
功能
容器后台线程,只有设置backgroundProcessorDelay大于0的容器才会启动ContainerBackgroundProcessor线程。该线程会调用当前容器的backgroundProcess方法,并且递归调用 backgroundProcessorDelay值小于等于0的子容器的方法。
从源码中看到只有StandardEngine设置了这个backgroundProcessorDelay值为10,所以只有StandardEngine容器启动ContainerBackgroundProcessor线程, 而其它StandardHost, StandardContext设置的值都是-1。
1 | /** |
2 | |
3 | * 创建一个新的StandardEngine组件,并绑定默认的基础Valve。 |
4 | |
5 | */ |
6 | |
7 | public StandardEngine() { |
8 | |
9 | super(); |
10 | |
11 | pipeline.setBasic(new StandardEngineValve()); |
12 | |
13 | /* Set the jmvRoute using the system property jvmRoute */ |
14 | |
15 | try { |
16 | |
17 | setJvmRoute(System. getProperty("jvmRoute")); |
18 | |
19 | } catch(Exception ex) { |
20 | |
21 | } |
22 | |
23 | // Engine将拥有reloading线程 |
24 | |
25 | backgroundProcessorDelay = 10; |
26 | |
27 | } |
线程类:ContainerBase.ContainerBackgroundProcessor
1 | /* |
2 | |
3 | * ContainerBase的保护线程类,调用当前容器的backgroundProcess方法,并在一个固定延时后, |
4 | |
5 | * 用它的子容器的backgroundProcess方法 |
6 | |
7 | */ |
8 | |
9 | protected class ContainerBackgroundProcessor implements Runnable { |
10 | |
11 | public void run() { |
12 | |
13 | while (!threadDone ) { |
14 | |
15 | try { |
16 | |
17 | Thread. sleep(backgroundProcessorDelay * 1000L); |
18 | |
19 | } catch (InterruptedException e) { |
20 | |
21 | ; |
22 | |
23 | } |
24 | |
25 | if (!threadDone ) { |
26 | |
27 | //获得当前容器,作为父容器 |
28 | |
29 | Container parent = (Container) getMappingObject(); |
30 | |
31 | ClassLoader cl = |
32 | |
33 | Thread. currentThread().getContextClassLoader(); |
34 | |
35 | if (parent.getLoader() != null) { |
36 | |
37 | cl = parent.getLoader().getClassLoader(); |
38 | |
39 | } |
40 | |
41 | //处理父容器和所有的子容器 |
42 | |
43 | processChildren(parent, cl); |
44 | |
45 | } |
46 | |
47 | } |
48 | |
49 | } |
50 | |
51 | //处理父容器和所有的子容器 |
52 | |
53 | protected void processChildren(Container container, ClassLoader cl) { |
54 | |
55 | try { |
56 | |
57 | //如果父容器的loader不为null,则将当前线程的上下文类加载器contextClassLoader设置为父容器 |
58 | |
59 | //的loader的类加载器 |
60 | |
61 | if (container.getLoader() != null) { |
62 | |
63 | Thread. currentThread().setContextClassLoader |
64 | |
65 | (container.getLoader().getClassLoader()); |
66 | |
67 | } |
68 | |
69 | //调用父容器的backgroundProcess方法 |
70 | |
71 | container.backgroundProcess(); |
72 | |
73 | } catch (Throwable t) { |
74 | |
75 | log.error("Exception invoking periodic operation: " , t); |
76 | |
77 | } finally { |
78 | |
79 | Thread. currentThread().setContextClassLoader(cl); |
80 | |
81 | } |
82 | |
83 | //获得父容器的所有子容器 |
84 | |
85 | Container[] children = container.findChildren(); |
86 | |
87 | for (int i = 0; i < children.length; i++) { |
88 | |
89 | //如果子容器的backgroundProcessorDelay小于等于0,则递归处理子容器 |
90 | |
91 | if (children[i].getBackgroundProcessorDelay() <= 0) { |
92 | |
93 | processChildren(children[i], cl); |
94 | |
95 | } |
96 | |
97 | } |
98 | |
99 | } |
100 | |
101 | } |
所属线程池
无
排查结果:
仔细看了栈,发现就是可以找到那个类名,就是一个线程。控制层和业务层怎么会不是一个线程。。。错误原因是我限制了for循环的个数,所以没找到。。。